John t whelan ranking simulator

[goblue78]

Re: John t whelan ranking simulator

It should make it impossible for a team to theoretically gain by losing.

Just to point out that it will always be possible (if not necessarily likely) to gain by losing. To take a simple example, suppose in the last game of the regular season a win gives you a bye in the first round of the playoffs while a loss makes you play a first round series. And assume you don't win the conference championship and get an autobid. By losing, you have the chance to go 5-1 in the playoffs (I'm thinking of the ECAC here, but other conferences work similarly) but by winning you can at best go 3-1 because of the first round bye. Those extra two wins might more than compensate for the last game loss. (It was even worse when the league had a third-place game, but no league does that now). The basic point is that the number of games you can win may be favorably affected by losing a game.

It would also be possible to lose a late game to boost your opponent into QWB bonus territory. You might then get points from earlier games against them in their rink. Even if it doesn't compensate for the loss, you might also knock someone else out of the QWB pool whom your opponent for a spot beat a couple of times on the road. So theoretically, at least, this system doesn't force you to always do better by winning. And it certainly doesn't do so if you're worried about seeding in the NCAAs, not just whether or not you get in.

[FlagDUDE08]

Re: John t whelan ranking simulator

Just to point out that it will always be possible (if not necessarily likely) to gain by losing. To take a simple example, suppose in the last game of the regular season a win gives you a bye in the first round of the playoffs while a loss makes you play a first round series. And assume you don't win the conference championship and get an autobid. By losing, you have the chance to go 5-1 in the playoffs (I'm thinking of the ECAC here, but other conferences work similarly) but by winning you can at best go 3-1 because of the first round bye. Those extra two wins might more than compensate for the last game loss. (It was even worse when the league had a third-place game, but no league does that now). The basic point is that the number of games you can win may be favorably affected by losing a game.

It would also be possible to lose a late game to boost your opponent into QWB bonus territory. You might then get points from earlier games against them in their rink. Even if it doesn't compensate for the loss, you might also knock someone else out of the QWB pool whom your opponent for a spot beat a couple of times on the road. So theoretically, at least, this system doesn't force you to always do better by winning. And it certainly doesn't do so if you're worried about seeding in the NCAAs, not just whether or not you get in.

And sometimes, it is best to go down to a weaker opponent than to move on and face an opponent that will severely weaken your chances, such as when RPI lost to Colgate in 2011; they would have had to face Cornell had they won.

[Patman]

The .850 vs. .851 sounds like rounding vs. truncation in terms of display. The application stores the full decimal, but only outputs the first three places; I assume Java truncates instead of rounds. Also, when you are calculating the Quality Wins Bonus, are you taking into account the games that were removed in negative effect on the RatingsPI? One new feature with the 0.4 version is it left in a debug statement that when you select a team, it will tell you what games have been removed in the command line. Obviously you'll need to run from the command line instead of the .exe in order to get this.

I'd be careful about blaming rounding... If an error is buried enough it can look like precision error vs human error

[Patman]

The cool thing would be to ask Knowlton and see. If he says "Sorry, I can't divulge that." then fine. Otherwise, if he says, then we know for sure that the removed games either are or are not counted in the divisor for the QWB bonus.

I will work up this KRACH Spreadsheet for you. It will take me a little time, though. I just checked, and I still have the spreadsheet I made earlier. If you have looked at any explanations, you know that KRACH requires iteration, and in my Excel sheet, you have to do that manually. Btw, I only have Excel2003. Like I say, this is a hobby, because I like Numbers.

In general, I like both the home/away weighting idea and the QWB that grades down through the RPI listing - with no TUC component. It should make it impossible for a team to theoretically gain by losing.

Spreadsheet? I may as well start drinking now.

Edit: I will get around to this during my Christmas break. Home should be boring enough to make appropriate changes. If FD is the new easy way for a game record I'll lean on him.

Edit #2: Dahl, I assume the game value is still relevant towards deletion... This stuff makes it needlessly complicated and as usual I like running Monte Carlo runs.

FD, how are you computing?

[FlagDUDE08]

Re: John t whelan ranking simulator

I'd be careful about blaming rounding... If an error is buried enough it can look like precision error vs human error

I'm well aware of this. However the winning percentage calculation, which we were discussing, is not complex enough for this to be the case. If we were talking about something such as OWP or OOWP, also the quality wins bonus, I am quite sure that this could be a programming or calculation error.

[FlagDUDE08]

Re: John t whelan ranking simulator

Spreadsheet? I may as well start drinking now.

Edit: I will get around to this during my Christmas break. Home should be boring enough to make appropriate changes. If FD is the new easy way for a game record I'll lean on him.

Edit #2: Dahl, I assume the game value is still relevant towards deletion... This stuff makes it needlessly complicated and as usual I like running Monte Carlo runs.

FD, how are you computing?

I wrote a Java application with various methods to do each portion of the calculation.

[Patman]

I'm well aware of this. However the winning percentage calculation, which we were discussing, is not complex enough for this to be the case. If we were talking about something such as OWP or OOWP, also the quality wins bonus, I am quite sure that this could be a programming or calculation error.

I wasn't reading closely . to me round at 3 places is round at 3 places

[Numbers]

Patman, I am very aware that a spreadsheet is slow. I have no access to anything faster, and no skill. I offered because Flag seemed unsure how to calculate, and a spreadsheet for KRACH would be instructive in some ways. As you know, KRACH calculates differently. I think it is an easier calculation than all the minutiae involved in the current pairwise.

If you want to work up a KRACH calculator, too great!!!

[Patman]

Patman, I am very aware that a spreadsheet is slow. I have no access to anything faster, and no skill. I offered because Flag seemed unsure how to calculate, and a spreadsheet for KRACH would be instructive in some ways. As you know, KRACH calculates differently. I think it is an easier calculation than all the minutiae involved in the current pairwise.

If you want to work up a KRACH calculator, too great!!!

It's not that it's slow, it's that it's painful

[FlagDUDE08]

Re: John t whelan ranking simulator

It's not that it's slow, it's that it's painful

Especially once you start getting into macros.

[goblue78]

Re: John t whelan ranking simulator

This is all interesting, and thanks, Flagdude, for your service here, but I'm spiritually with patman. Sometime around New Years' I'll reprogram my previous lightning fast Stata/Mata code and do this for real. (It also does KRACH and home/road/tie adjusted KRACH.) But I like the Java App because at least it will tell me if I'm matching your results. As I've done before, I'm happy to make that code available to everyone similarly obsessed. And making the results available on a dynamic basis is a real service, FD. Thanks.

[Numbers]

Re: John t whelan ranking simulator

This is all interesting, and thanks, Flagdude, for your service here, but I'm spiritually with patman. Sometime around New Years' I'll reprogram my previous lightning fast Stata/Mata code and do this for real. (It also does KRACH and home/road/tie adjusted KRACH.) But I like the Java App because at least it will tell me if I'm matching your results. As I've done before, I'm happy to make that code available to everyone similarly obsessed. And making the results available on a dynamic basis is a real service, FD. Thanks.

If you are working on an app that does KRACH I am not going to bother.

Thanks everyone for being at least as obsessed as I am.

By the way - I emailed the Minnesota rep on the committee this morning about the detail in the QWB calculation we are wondering about, and have not heard any reply yet.

[Patman]

This is all interesting, and thanks, Flagdude, for your service here, but I'm spiritually with patman. Sometime around New Years' I'll reprogram my previous lightning fast Stata/Mata code and do this for real. (It also does KRACH and home/road/tie adjusted KRACH.) But I like the Java App because at least it will tell me if I'm matching your results. As I've done before, I'm happy to make that code available to everyone similarly obsessed. And making the results available on a dynamic basis is a real service, FD. Thanks.

I still say the end goal is a simulator

Edit: I am not using a modular executable language... I don't know the differences in the -oriented but what I do is use a thing that primarily uses C as a platform. I suppose its possible to treat it as a script but not without installing software.

For me KRACH is deadly simple once you purée the data into a win matrix and game matrix. I've posted the code for that before.

I'll say the big thing is if we can adopt a data input standard that will go a long way.

[Numbers]

Re: John t whelan ranking simulator

Flag, Jim Dahl,

Just a note. Your RPI tables do not agree with each other, and neither agrees with what is posted on USCHO or CHN (those 2 have the same).

[FlagDUDE08]

Re: John t whelan ranking simulator

Flag, Jim Dahl,

Just a note. Your RPI tables do not agree with each other, and neither agrees with what is posted on USCHO or CHN (those 2 have the same).

USCHO and CHN aren't going to agree; neither of those sites have correctly taken into account the new calculation rules, and I know USCHO will not until January. As for Jim's and my differences, we've already discussed a disagreement in Quality Wins Bonus. I will also ask Jim if he is taking weighting into account on RatingsPI, because it shouldn't be happening. One other factor that could be making a difference is how OOWP is calculated, specifically whether games involving the team in question should be counted. Some sources say yes, others say no. When I was using USCHO as a model very early in the season with teams that had a RatingsPI that was easy to calculate, I found the answer to be no.

One other thing that could cause issue is specifically how OWP and OOWP is calculated. Do you take a cumulative record, or do you take the average of each team's records?